Vitality’s AI‘s Emile Stipp explains how integrating artificial intelligence (AI), behavioural science, and trusted data is reshaping insurance, risk management and client engagement.
At the recent launch of Vitality AI’s partnership with Google, held at London’s prestigious Langham Hotel, leaders from insurance, technology, and academia gathered to discuss how AI can reshape global health outcomes. The collaboration, uniting Discovery Group and Google, aims to accelerate early disease detection and empower individuals to make meaningful health changes through personalised data insights.
Among the speakers was Emile Stipp, Managing Director of Vitality AI and Global Chief Actuary at Discovery. This exclusive conversation offered a compelling look into how AI is redefining insurance, both as an efficiency tool and as a model for moral and behavioural transformation.
Beyond efficiency: AI as a business model driver
Many organisations approach AI through a narrow operational lens for automating processes, cutting costs, or boosting staff productivity. Stipp sees things differently.
“For us, it’s really about building AI into the actual business model,” he said. “It’s about driving change that’s different from just doing your accounts more efficiently. It has the almost moral purpose of making people healthier.”
That moral dimension, he explains, is central to how Vitality has integrated data science across its insurance products. Rather than simply deploying algorithms to streamline underwriting or claims, Vitality AI aims to use intelligence as a behavioural catalyst, enabling proactive, personalised health interventions that weren’t possible even five years ago.
“We couldn’t really talk to people in a personalised way about their health because the technology wasn’t ready,” Stipp noted. “Now we can, and if you point AI at the right target variable, you get much more meaningful answers.”
The intention is clear: to make every product, policy, and interaction part of a single data-driven ecosystem that connects lifestyle choices, health outcomes, and financial protection.
Building trust through data integrity
Behind the sophistication of AI models lies something less glamorous but essential: data quality. As Stipp put it, “It’s taken us more than two decades to make sure we have quality data.”
Since the late 1990s, Vitality has been investing in data architecture and standardisation, a commitment that predates the AI boom but now forms the backbone of its global technology strategy.
“We apply what we call medallion structures, a system that rates how trustworthy the data is, from bronze to silver to gold,” Stipp explained. “Only gold-level data goes into our models. That means it’s cross-referenced, the lineage is clear, and we understand exactly why it says what it does.”
This depth of verification, he says, prevents common pitfalls that have derailed other AI initiatives, such as inaccurate records, unverified health claims and inconsistencies in diagnosis data. “It’s essential. If you don’t do it, you’re in trouble,” he said.
Data accuracy is not only an operational concern but also a matter of ethics and transparency. Each AI-driven recommendation within Vitality’s systems is traceable and explainable.
“Whenever a model tells you something, you can click through to see why,” Stipp said. “We’ve built explainability into the system for over a decade now. That transparency means people trust what they see, and interpret it correctly.”
Personalisation and prevention in practice
The real impact of AI, according to Stipp, comes from its ability to make relevance scalable. The technology translates data into actionable insights that resonate with individual users, whether warning a customer about grapefruit interactions with statins or promoting better sleep after tracking behavioural patterns.
“It’s so relevant to people that it becomes powerful,” he said. “We expected we’d have to work really hard to drive engagement in the first year, but even after three to six months, engagement was more than ten times what we expected.”
Personalisation is also enabling Vitality to reach typically unengaged populations. “In some markets, WhatsApp is the key channel,” he explained. “In others, like the US, it’s iMessage. The challenge is to find the right way to reach people wherever they are.”
This localisation extends to the company’s global engagement goals. Targets differ by market, with Japan’s already-active population requiring different incentives from regions where sedentary lifestyles are more common.
“Everybody should always have a goal,” Stipp said. “We set targets for screenings, for exercise, for sleep, but they have to be contextual. The first phase is activation, then we learn and adapt.”
Vitality aims to raise cancer screening participation from 30% to two-thirds within a year. Even if targets aren’t fully achieved, Stipp said, “We learn something every time, about how to make the programme better.”
Future responsiveness: learning from data at scale
If the last few years have proved anything, it’s that adaptability is critical. The COVID-19 pandemic highlighted the link between lifestyle and resilience, insights that Stipp’s team observe directly through Vitality’s data ecosystem.
“People who exercised had better outcomes, even if they had COVID,” he recalled. “Lifestyle is just so fundamental to human health that no matter what happens, it affects everything.”
Because Vitality’s AI systems are built on feedback loops, they can adjust dynamically as new patterns emerge. “The models learn, we learn from the models, and they see how people react,” said Stipp. “That makes the whole system more responsive.”
Responsiveness extends to employers, too. Corporate clients can receive anonymised insights on workforce wellbeing, anything from mental health patterns to engagement trends across divisions. “You can literally point it at one person or at an entire organisation,” Stipp explained. “You can say, ‘show me BMW’s experience,’ and it’ll generate a tailored report on what’s happening.”
He stressed that privacy is paramount. “The smaller the group, the less specific we can be. But at scale, it gives incredibly valuable insights that help employers support their people better.”
Towards a healthier system
The intersection of AI, behaviour, and health economics represents the next frontier of insurance innovation. Tools such as GLP-1 weight-loss drugs are reshaping how people manage health, but they must be paired with behavioural change to sustain impact.
“We see from the data that people get better outcomes if they start losing weight and take up exercise at the same time,” he said. “That’s the message we want to get across: making the drug payment contingent on maintaining healthy activity. “The very definition of habit is that it becomes self-motivating. Once it’s a habit, it’s not that hard to do it again tomorrow.”
AI, in Vitality’s model, is not just about prediction but empowerment. It bridges the gap between data and human decision-making, encouraging smarter choices, better outcomes, and stronger trust in the system.
“If we get this right, we can solve so much, from economic pressures to NHS strain,” Stipp concluded. “When people lose weight, exercise, and live healthier lives, it changes everything.”
